Cloud Bowl

A game where microservices battle each other in a giant real-time bowl.

Run Locally:

1. Make sure you have docker installed and running
2. Start Kafka

   ./sbt "runMain apps.dev.KafkaApp"
   

3. Start the sample service

   (cd samples/scala-play; ./sbt run)
   

4. Start the Battle

   ./sbt "runMain apps.Battle"
   

5. Start the apps.dev Kafka event viewer

   ./sbt "runMain apps.dev.KafkaConsumerApp"
   

6. Start the apps.dev Kafka event producer

   ./sbt "runMain apps.dev.KafkaProducerApp"
   

   You can send commands like

   ARENA/create
   ARENA/unlist
   ARENA/playerjoin
   ARENA/playerfill/NUM
   ARENA/playerleave/ID
   ARENA/viewerjoin
   ARENA/scoresreset
   

7. Start the Viewer web app

   ./sbt run
   

   Check out the foo arena: http://localhost:9000/foo

Web UI Notes:

Pause the Arena refresh:

document.body.dataset.paused = true;

Run on Google Cloud

1. Create GKE Cluster with Cloud Run

   gcloud config set core/project YOUR_PROJECT
   
   gcloud services enable compute.googleapis.com container.googleapis.comcontainer.googleapis.com containerregistry.googleapis.com cloudbuild.googleapis.com
   
   gcloud config set compute/region us-central1
   gcloud config set container/cluster cloudbowl
   
   gcloud container clusters create \
     --region=$(gcloud config get-value compute/region) \
     --addons=HorizontalPodAutoscaling,HttpLoadBalancing,CloudRun \
     --machine-type=n1-standard-4 \
     --enable-stackdriver-kubernetes \
     --enable-ip-alias \
     --enable-autoscaling --num-nodes=1 --min-nodes=0 --max-nodes=20 \
     --release-channel=stable \
     --scopes cloud-platform \
     $(gcloud config get-value container/cluster)
   

2. Install Strimzi Kafka Operator

   kubectl create namespace kafka
   curl -L https://github.com/strimzi/strimzi-kafka-operator/releases/download/0.20.0/strimzi-cluster-operator-0.20.0.yaml \
     | sed 's/namespace: .*/namespace: kafka/' \
     | kubectl apply -f - -n kafka
   

3. Setup the Kafka Cluster

   kubectl apply -n kafka -f .infra/kafka.yaml
   kubectl wait -n kafka kafka/cloudbowl --for=condition=Ready --timeout=300s
   

4. Get your IP Address:

   export IP_ADDRESS=$(kubectl get svc istio-ingress -n gke-system -o 'jsonpath={.status.loadBalancer.ingress[0].ip}')
   echo $IP_ADDRESS
   

5. Create a ConfigMap named cloudbowl-config:

   export APPLICATION_SECRET=$(head -c 32 /dev/urandom | base64)
   export ADMIN_PASSWORD=PICK_A_PASSWORD # Used for creating / updating arenas
   cat <<EOF | kubectl apply -f -
   apiVersion: v1
   kind: ConfigMap
   metadata:
     name: cloudbowl-config
   data:
     WEBJARS_USE_CDN: 'true'
     APPLICATION_SECRET: $APPLICATION_SECRET
     ADMIN_PASSWORD: $ADMIN_PASSWORD
   EOF
   

6. Setup Cloud Build with a trigger on master, excluding samples/**, with Configuration Type set to Cloud Build configuration file, and substitution vars _CLOUDSDK_COMPUTE_REGION and _CLOUDSDK_CONTAINER_CLUSTER. Running the trigger will create the Kafka topics, deploy the battle service, and the web app.

7. Once the service is deployed, setup the domain name:

   export IP_ADDRESS=$(kubectl get svc istio-ingress -n gke-system -o 'jsonpath={.status.loadBalancer.ingress[0].ip}')
   echo "IP_ADDRESS=$IP_ADDRESS"
   
   gcloud beta run domain-mappings create --service cloudbowl-web --domain $IP_ADDRESS.nip.io --platform=gke --project=$(gcloud config get-value core/project) \
     --cluster=$(gcloud config get-value container/cluster) --cluster-location=$(gcloud config get-value compute/region)
   gcloud compute addresses create cloudbowl-ip --addresses=$IP_ADDRESS --region=$(gcloud config get-value compute/region)
   

8. Turn on TLS support:

   kubectl patch cm config-domainmapping -n knative-serving -p '{"data":{"autoTLS":"Enabled"}}'
   kubectl get kcert
   

Troubleshooting

# Pick a topic:
export TOPIC=viewer-ping
export TOPIC=players-refresh
export TOPIC=arena-update

# Describe a topic:
kubectl -n kafka run kafka-consumer -ti --image=strimzi/kafka:0.17.0-kafka-2.4.0 --rm=true --restart=Never -- bin/kafka-topics.sh --describe --bootstrap-server cloudbowl-kafka-bootstrap.kafka:9092 --topic $TOPIC

# Consume messages on a topic:
kubectl -n kafka run kafka-consumer -ti --image=strimzi/kafka:0.17.0-kafka-2.4.0 --rm=true --restart=Never -- bin/kafka-console-consumer.sh --bootstrap-server cloudbowl-kafka-bootstrap.kafka:9092 --topic $TOPIC --from-beginning --property print.key=true --property key.separator=":"